Analysis of decay heat in spent nuclear fuel using serpent 2 simulations and Clab measurements

This study compares measured and calculated decay heat values for spent nuclear fuel assemblies stored at Sweden’s Clab facility, an interim storage site where decay heat can be directly measured. The analysis utilizes measured values from two reports: one from the Svensk Kärnbränslehantering Aktiebolag (SKB, 2006) and another from the Electric Power Research Institute (EPRI, 2024). A number of these measured values is present in both reports, possibly updated in the newer version. Burnup and decay calculations are performed using the Monte Carlo software Serpent 2 to accurately model the assemblies’ geometry and the respective reactors’ operating conditions. The impact of using two different sources of information on decay heat measurements is assessed through C/E (Calculations over Experiments) ratios for the selected fuel assemblies. The results indicate no statistical evidence that the re-evaluated estimates from the EPRI report differ significantly from the initial estimates in the SKB report. Furthermore, the study highlights that even small variations in moderator density can lead to significant changes in the final decay heat values. Finally, the study evaluates differences in C/E ratios across various fuel assembly types. Confidence intervals are estimated first at a general level, based on the reactor type, and then refined to account for specific fuel assembly characteristics. In general we found the 95% confidence interval for the C/E of a fuel assembly taken from EPRI report to range from 1.016 to 1.023, with a mean C/E of 1.019. • Comparison between measurements from Clab facility and Serpent 2 simulations. • A total of 270 measurements for EPRI report were compared to Serpent 2 simulations. • A total of 109 measurements for SKB were compared to Serpent 2 simulations. • Two different nuclear data libraries were used (ENDF/B-VII.0 and ENDF/B-VII.1). • The mean C/E is 1.019 for the EPRI report data, and 1.011 for the SKB data.